Journal ofCoastal Research 1151-1156 West Palm Beach, Florida Fall 2000

Pollution by Mud of Great Barrier Reef Coastal Waters Eric Wolanski and Simon Spagnol

Australian Institute of Marine Science PMB No.3 Townsville Me, Qld. 4810, Australia ABSTRACT _

WOLANSKI, E. and SPAGNOL, S., 2000. Pollution by mud of Great Barrier Reef coastal waters. Journal ofCoastal Research, 16(4), 1151-1156. West Palm Beach (Florida), ISSN 0749-0208. Oceanographic data from the coastal waters of the Great Barrier Reef near Cairns were obtained for two months in the 1997 dry season when river runoff was negligible. A muddy coastal zone was found with suspended sediment concentration reaching 1000 mg/L during trade winds. These high turbidity events are very stressful conditionsfor coral reefs and were caused by resuspension of sediment by wind waves. In calm weather a nepheloidlayer prevailed, extending about 30 km offshore and carrying coastal mud toward offshore coral reefs. In this layer, muddy marine snow with floes exceeding 1000 urn in diameter was found inshore, while only small floes < 100 urn were found offshore. Near-surface visibilitymay have decreased by 50% in the last 70 years. High turbidity appears to be due to man and ultimately threatens the main bodyof the Great Barrier Reef.

ADDITIONAL INDEX WORDS: Mud, turbidity, siltation, resuspension, nepheloid layer, Great Barrier Reef.

INTRODUCTION METHODS

Most of our knowledge of the historical, physical properties For six weeks, in August and September 1997, six ocean­ of the waters ofthe Great Barrier reef (GBR) is derived from ographic moorings were maintained in a cross-shore transect the 1928-1929 British Museum Expedition. This pioneering at sites 1-6 shown in Figure 1 in depth varying from 3 m study was mostly carried out near Low Isles (16°23'S, inshore to 23 m offshore. Water depth and equipment de­ 145°34'E; see Figure 1 for a location map). MOORHOUSE ployed are summarised in Table 1. At sites 1-5 the moorings (1933) and ORR (1933 a and b) reported clear water (mean included Inter-Ocean model S4 current meters at about half­ visibility =11 m) during the South East trade wind season depth, a Dataflow salinometer and a total of 19 nephelome­ ters spread across the water column, the lowest one being from about April to October. Human impacts at that time located 0.5 m from the bottom. The units averaged data over were negligible. Since that time most of the forest and nat­ 1 min and recorded these data at 5 min intervals. ural vegetation that covered the land have been cleared and At site 6 near Pixie Reef an Inter-Ocean model S4 current much of it is now intensively farmed with usually no protec­ meter was deployed 6m from the bottom, and a Dataflow sa­ tive cover against soil erosion. This resulted in a large in­ linometer at 17m off the bottom in 25 m depth. During moor­ crease in mud discharged by rivers on the coast of the Great ing deployment and recovery, and also in December 1997 Barrier Reef (ANON, 1993; WOLANSKI, 1994, LARCOMBE et when dry weather conditions prevailed, vertical profiles of al., 1996, WACHENFELD et al., 1997). Historical photographs salinity, temperature and suspended solid concentration and navigation maps of the Cairns waterfront (see location (SSC) were obtained using a CTD equipped with an Analite in Figure 1) suggest that what was largely a sandy coast is backscatterance nephelometer. The nephelometers were cal­ now buried under 1.5 m of mud (WOLANSKI and DUKE,2000). ibrated in-situ for suspended solid concentration. In-situ pho­ The increased muddiness of the coast has been reported also tographs of the suspended matter were obtained using the at other locations along the coast of the Great Barrier Reef method of AYUKAI and WOLANSKI (1997). following European settlement. For instance, WOLANSKI Suspended solids were collected at 2.5 m from the bottom (1994) reported an invasion by mud of a sandy beach flat at every 2 days at site 5 using a McLane automated suspended Townsville (19.3°S). Large areas of mangrove forests that act­ sediment trap and the samples were was analyzed in the lab­ ed as efficient sediment traps (FURUKAWA et al., 1997), have oratory following the techniques described by WOLANSKI et also been removed by farming. Many estuaries essentially ai. (1999). have become drains bringing all the eroded, fine sediment Wind and rainfall data were obtained from meteorological directly to the sea. In view of this man-made increased mud­ stations at Green Island and Cairns. diness of coastal waters, oceanographic data were collected to assess the offshore extent of the dispersion of the mud and RESULTS the potential threat it poses to the Great Barrier Reef. Southeast trade winds peaking at 15 m S-1 prevailed throughout the study period, occasionally interrupted by 99050 received 17 May 1999; accepted in revision 22 December 1999. short periods of calm weather (Figure 2). The tides were 1152 Wolanski and Spagnol

Table 1. Mooring details.

,~. '·'i~'·\,··f Depth of ~ Site Water current Depth of Low Isles ;~BallRf. number depth (m) meter (m) nephelometer (rn) 1 2 N/A 0.25,1 2 4 2.5 1,2,3,3.5 Pixie Rf. 3 6 4.5 3,4,5,5.5 6 '~ 4 11 6 6, 8, 10, 10.5 -f 5 18 12 14, 17, 17.5 6 25 19 N/A

made of mucus, plankton detritus as well as other biological detritus. Offshore the suspended solids were present as in­ organic mud floes < 60 J.Lm in diameter. The spectrum of sea level (not shown) revealed that the semi-diurnal tides contained most of the energy. On the other hand the alongshore currents and wind showed negligible en­ ergy at diurnal and semi-diurnal tidal frequencies; most of the energy was contained at low-frequencies with period> 3 days with no apparent peak in the spectra. The wind showed a weak peak at diurnal frequency, an indication of a sea­ Figure 1. General location map showing the mooring sites. breeze effect. The spectra of SSC at stations 1 and 2 were weakly energetic at the diurnal frequency, possibly an effect of the sea-breeze, but not at all at the dominant tidal semi­ mostly semi-diurnal with a strong diurnal inequality, peak­ diurnal frequency. Highest energy in the spectra was at ing at about 3 m peak-to-trough at spring tides and 0.5 m at about 10 days period. This period is the same for the peak in neap tides (Figure 2). Freshwater runoff and associated riv­ the wind and current spectra, it is also the maximum period erine sediment inflow was negligible during the study be­ for which the spectrum can be reliably calculated from this cause it was carried out in the dry season. time series. The currents were mostly alongshore, alternating north­ Alongshore wind offshore (at Green Island) and currents ward and southward at periods of several days, the tidal var­ were significantly correlated with a lag increasing with dis­ iations were small and the low-frequency currents were dom­ tance offshore, the wind preceding the current (R = 0.8 at 2 inant (Figure 2). The currents were southward with speed of h lag at site 2; R = 0.6 at 8 h lag for site 3 and R =0.8 at 16 about 0.2 m S-l during calm weather and northward with a h lag at sites 5 and 6). speed of about 0.4 m S-l during strong trade winds and fluc­ The alongshore currents and the SSC were also signifi­ tuated very little with the tides. cantly correlated, the current preceding SSC, R = 0,.7 at 2 The suspended solid concentration (SSC) fluctuated widely hr lag at site 2, R = 0.45 at 6 hr lag at site 3, R = 0.5 at 13 in coastal waters (Figure 2). Maximum SSC was about 1,000 h lag at site 4. A negative and significant correlation between mg 1-\ at which time the visibility as reported by divers at wind and current occurred at site 5 located further offshore, site 1 was zero (it was pitch black at 2 m depth and the divers R = -0.5 at zero lag. No correlation (lRI < 0.2) between cur­ could not see their hand against the face mask). The SSC was rents and SSC prevailed at site 5. smaller further offshore. The SSC did not vary with the tides. For sites 1 and 2 a significant correlation also existed be­ Inshore, high SSC occurred frequently, but only during tween alongshore wind velocity and SSC; at site 1 R = 0.7 strong wind; at these times SSC was uniform throughout the at zero lag and 0.9 at 14 h lag; at site 2 R = 0.7 at 2 h lag. water column. Offshore, high SSC events (>50 mg 1-1 peak­ For site 3, no strong correlation was apparent (R < 0.4). At ing in one event at 200 mg 1-1) occurred occasionally. At these site 4 the correlation was negative and significant, R = -0.45 times the high SSC events were limited to near the bottom at zero lag. At site 5 no correlation was found between the (Figure 3a). During strong winds, suspended solids consisted wind velocity and the SSC (lRI < 0.2). of fine terrigenous mud forming floes typically 30-60 J.Lm in The SSC fluctuated coherently with distance offshore be­ diameter (not shown). tween sites 1-4. Between sites 1 and 2, R = 0.75 at 0.4 h lag; In calm weather a near-bottom nepheloid layer was ob­ between sites 1 and 3, R = 0.73 at 2.3 h lag and between served apparently cascading down the slope (Figure 3b). At sites 1 and 4, R = 0.65 at 7.3 h lag. The correlation however 1 that time in coastal water SSC reached 50 mg 1- • Offshore became negative but was still significant for sites further off­ 1 SSC values in the nepheloid layer peaked at about 10 mg 1- . shore, e.g. between sites 1 and 5 R = -0.5 at 16 h lag. Inshore, the suspended matter consisted of muddy marine The suspended solids collected every 2 days at site 5 in a snow, i.e. small mud floes (typically 30-80 J.Lm in diameter) McLane sediment trap were largely inorganic. TOC varied attached to sticky marine snow floes typically 200-1000 J.Lm from sample to sample between 1.4 and 3.7%, TIC between in diameter (Figure 3c). The marine snow was observed to be 1.0 and 3.7%, Ca between 6.6 and 8.5%, Si between 16.3 and

Journal of Coastal Research, Vol. 16, No.4, 2000 Mud Pollution at Great Barrier Reef 1153

Wind 15 Green Island

I --en 0 E

-15 Sea Level 8

E 6

4 Currents 0.5

I --en 0.0 E

-0.5

1.0 SSC

Site 4 C') I E 0.5 0> ~

0.0

1.0

Site 2 C') I E 0.5 0> ~

0.0

1.0 Site 1 C') I E 0.5 C) ~

0.0

1-Aug-97 8-Aug-97 15-Aug-97 22-Aug-97 29-Aug-97 5-Sep-97 12-Sep-97

Figure 2. Time series plot of the wind vectors (shown using the oceanographic convention), sea level, currents, and examples of the measured near­ bottom suspended solid concentration.

Journal of Coastal Research, Vol. 16, No.4, 2000 1154 Wolan ski and Spagno l

1050h 19Aug 1997

(a)

30 35

29 Dec 1998 (b) sse

35

(c)

Figure 3. (a) shows the cross -shore dist ribution of SSC during strong wind s at 1050 h on August 19, 1998, as measured by the neph elometer s on the 1 moorin gs. At that tim e only small floes (diameter < 60 mg 1- ) wer e present at all sites . (b) shows the cross-shore distribution on December 29, 1998, still in the dry season, in calm weather conditions . Note the near-bottom neph eloid layer. (c) Micro-photographs spa nning 250 urn in width, also taken on December 29, 1998, showing th e presen ce of mudd y marine snow in inshore wat ers, and of small floes of diamet er < 60 urn in offshore wat ers.

20.1 and Al betw een 7.5 and 8.3. The sus pended sediment lag between wind and currents increasing with water depth was thus mainly terrigenous mud. or distan ce offshore. What is new is the findin g of high values (up to 1,000 mg DISCUSSION I - 1) of sus pended solid concentration (SSC) in coastal waters. Also new is the finding of a nepheloid layer with fine sedi­ The water circulation in this area of the Great Barrier Reef ment apparently cascading from inshore to offshore in calm was simila r to th at previously reported by WOLANSKI(1994). weath er. The high SSC occurred only durin g st rong wind s In particular the macro-tid es (up to 3 m peak-to-trough ) prop­ and in inshore waters. It was mad e visible by a band of mud­ agated perp end icular to the coast , genera ting weak tidal cur­ dy water a few km wide parallel to the coast and extending rents in coastal wat ers. The dominan t current in calm weath­ northward and southwa rd as far as one could see from th e er was alongshore southward, dri ven by th e southward flow­ ship. Visual observations at sea und er st rong winds revealed ing East Australian Current on the continental slope of th e patches of muddy coastal waters moving offshore severa l km Great Barrier Reef. Strong southeasterly trad e wind s re­ from th e coast . Riverine sediment inflow was negligible dur­ versed the current from southward to northward, with th e ing that period, th e dry season. The materi al making the wa-

Journal of Coastal Research, Vol. 16, No. 4, 2000 Mud Pollution at Great Barrier Reef 1155 ters turbid in coastal waters was terrigenous mud res us­ pended by waves during strong winds, during strong winds 16.0 no marine flow formed and the flocs remai ned sma ll. On the outer edge of the coastal band, site 4, there was a negative correlation between wind an d SSC. This finding sug­ gests that the wind waves during the study period did not 12.0 resuspend sediment at that depth, and that the nepheloid Low Isles 1927 layer formed consistently in calm weather.Hence the coastal Ul mud is locally stirred by wind and currents inshore an d prop­ e Q; agates offshore in calm weathe r only. oS In calm weather muddy marine snow formed inshore. It :E 8.0 was similar to that described by AYUKAI an d WOLANSKI :e (1997) and WOLANSKI et (1999) for nu trient-rich, muddy .!!! at. > tropica l coastal waters elsewhere along the muddy tropical coast of the Great Barrier Reef. 4.0 During the shi p-born study of this process , calm weather prevailed. The neph eloid layer exte nded offshore past site 5. The currents at th e time were southward with speed < 0.2 m S - 1. This speed was too small to entrain the bottom mud but sufficient to generate near-bottom tu rbulence inhibiting 0.0 -+-"-"'- -,----- ,------,- - ,-- -,-- --f"''--- ,------, the settling on the bottom of suspen ded mud. WOLANSKI et 4 8 12 16 al. (1992) reported a similar example of dredged mud dump ed Distance from Coastline (km) at sea, settling towa rd th e bottom, prevented by near-bottom Figure 4. Mean water visibility near th e sur face and its standard devi­ turbulence to reach the bottom and, instead, forming a ne­ ation shown as error bars, at the mooring sites in August-September pheloid layer. Thus, we suggest the neph eloid layer was gen­ 1997. Note the increasing visibility with distance offshore. In 1997 at site erated inshore and propagated offshore . Bottom turbulence 5 located 12 km from th e coast th e mean visibility near th e surface was associated with the prevailing current was sufficiently large half that measured in 1927 by ORR (1933a) at Low Isles (see Figure 1 for a locati on map ). to maintain the near-bottom nepheloid layer in suspension and too small to entrain the sediment upward throughout the water column. This hypoth esis is consistent with the finding of a decrease in SSC in the nepheloid layer with distance Intense tropical rain in the wet season (December-March) offshore. It is also consistent with the negative corre lation gene rates severe erosion of cleared in this area (WOLANSKI, between currents and SSC at site 4, while the correlation was 1994). Increased muddiness of coastal water results . This is positive at the inshore sites 1-3. apparent along the Cairns waterfront which, as historical There exists a strong negative correlation (R2 = 0.8) be­ photographs and navigation charts revealed, is now a 1.5 m tween visibility (measured by a secci disk) and SSC (WOLAN­ thick mud layer covering what was a sandy beach 100 years SKI et al., 1981). Accordingly it was possible to compare the ago (WOLANSKI and DUKE , 2000). Presumably mud and nu­ near-surface visibility along the tran sect with th at measured trients from land runoff may be res ponsible for th e acute de­ by ORR (1933a) at Low Isles in 1927 also in the dry season. terioration of coral reefs at Double Island (near site 2) where (Figure 4). Low Isles is situated North of th e transect at the our underwater observations revealed that most (probably same distance from the shore as site 5. In 1927 human im­ > 90%) hard corals on the west side were dead , a few still had pacts on the Great Barrier Reef were probably smaller than a thin coating of mud in grooves, most were completely cov­ at present becau se far ming was less exte nsive though indis­ ered by algae, and soft corals were growing over hard corals criminate deforestation was underway (RATCLIFFE, 1947). at the remaining places. At Low Isles, BELL an d ELMETRI The wind direction and speed during August-Septe mber (1995) reported a shift from a pre dominance of hard corals in were similar in 1927 and 1997, comprising periods several 1927 to a predomina nce of soft corals at present. Similar days to several weeks in duration of strong southeasterly changes (ROGERS, 1983 and 1990; YAMAZATO, 1987; BELL,. trade winds, separated by periods of calm weather. Furthe r­ 1992; McCLANAHAN and OBURA, 1997) have been observed more there is negligible runoff during the dry season. Thus in othe r places where coral reefs have been subject to in­ it appears justified to compa re visibility in the dry seasons creased pulses of both nutrients an d mud. The mud contains of 1927 (at Low Isles) and 1997 (at site 5). most of the nutrients an d these are only released into the If the water visibility near Low Isles in 1927 was typical water column du ring wind stirring (WALKER and of historical conditions for this region of the Great Barrier O'DONNELL, 1981). The reaction of a pristine coral reef to Reef, we can only conclude that human activities already pulses of both nutrients and mud from land ru noff is presum­ have halved the visibi lity . If true, this is an alarming res ult ably always negative, that is enviro nmental degradation; re­ because it is a clear ind icat ion of a serious pollution of the covery, if any, can vary spatially and temporally as a resul t Great Barrier Reef by mud following land clearing, intensive of varying ecological processes (MCCOOK, 1994).The scien­ far ming, dredging, removal of mangroves and other human tific problem of separating natu ral from human impacts on activities increasing the discharge of mud in coastal waters. coral reefs is difficult in the pre sence of tropical cyclones that

J ournal of Coastal Researc h, Vol. 16, No.4, 2000 1156 Wolanski and Spagnol can occur in the wet season and may cause major geomor­ BELL, P.R.F., 1992. Eutrophication and coral reefs-some examples phologic changes to reefs (DONE, 1990). Low Isles coral reefs in the Great Barrier reef lagoon. Water Research, 26, 553-568. BELL,P.R. and ELMETRI, 1.,1995. Ecological indicators of large-scale for instance were damaged by tropical cyclones in 1934 and eutrophication in the Great Barrier Reef lagoon. Ambio, 24, 208­ 1950. However, only for the 1950 cyclone, when presumably 215. more mud was available for resuspension as a result of ero­ DONE,T.J., 1990. Effects of tropical cyclone waves on ecological and sion from cleared land, was mud explicitly mentioned as hav­ geomorphological structures on the Great Barrier Reef. Continen­ ing killed corals (HILL, 1985 a and b). tal ShelfResearch, 12, 859-872. HILL, D., 1985a. The Great Barrier ReefCommittee, 1922-1982: The If human-derived turbidity is already a problem in the dry first thirty years. Historical Records ofAustralian Science, 6 (1), season, it certainly is an even more serious problem during 1-18. the wet season and during tropical cyclones, for which no HILL, D., 1985b. The Great Barrier ReefCommittee, 1922-1982: The data are available. Last three decades. Historical Records ofAustralian Science, 6 (2), 195-221. The finding of mud cascading from inshore to offshore in a LARCOMBE, P.; WOOLFE, K.J., and PURDON, R.G., 1996. Terrigenous nepheloid layer in calm weather suggests that pulses of fine sediment fluxes and human impacts. CRC Reef Research Centre, terrigenous mud may also ultimately reach the main body of Current Research, Townsville, Australia, 174pp. the Great Barrier Reef in calm weather. It would then be MCCLANAHAN, T.R. and OBURA, D., 1997. Sedimentation effects on shallow coral communities in Kenya. Journal Experimental Biol­ brought to the surface during strong trade winds. There it is ogy Ecology, 209, 103-122 diluted by the fine calcareous sediment continuously pro­ MCCOOK, L.J., 1994. Understanding ecological community succes­ duced by pelagic forams and borers on the coral (WOLANSKI sion. Vegetatio, 110, 115-147. et al., 1999). Present estimates of the threat to the Great Bar­ MOORHOUSE, F.W., 1993. The temperature of the waters in the an­ rier Reef from terrigenous mud rely mainly on analysis of the chorage, Low Isles. British Museum (Natural History) Great Bar­ rier Reef Expedition 1928-29. Science Report Vol. II, pt. 4, 98-107. carbonaceous content of surface sediment. This assumption ORR, A.P., 1933a. Variations in some physical and chemical condi­ assumes a quasi-steady state situation and it neglects the tions on and near Low Isles Reef. British Museum (Natural His­ pulse-like intrusion of terrigenous mud. Except when the sys­ tory) Great Barrier Reef Expedition 1928-29. Science Report Vol. tem is overwhelmed by terrigenous mud, at which time it II, pt. 4, 87-98. ORR,A.P., 1933b. Physical and chemical conditions in the sea in the may be too late for remedial actions, the analysis of grab neighborhood of the Great Barrier Reef. British Museum (Natural samples is thus an inappropriate method. History) Great Barrier Reef Expedition 1928-29. Science Report Vol. II, pt. 3, 37-86. CONCLUSION RATCLIFFE, P., 1947. Flying fox and drifting sands. Angus & Rob­ ertson, Melbourne, 332 pp. We suspect that environmental degradation of the Great ROGERS, C., 1983. Sublethal and lethal effects of sediments applied Barrier Reef will continue unabated, indeed will worsen, as to common Caribbean reef corals in the field. Marine Pollution Bulletin, 14, 378-82. land clearing continues and farming activities intensify, sim­ ROGERS, C., 1990. Responses of coral reefs and reef organisms to ply because, in practice, little is done to prevent pollution by sedimentation. Journal Experimental Marine Biology Ecology, 62, mud notwithstanding the World Heritage status of the Great 185-202. Barrier Reef. The problem is not only social and economical, WACHENFELD, D.; OLIVER, J., and DAVIS, K., 1997. State ofthe Great Barrier Reef World Heritage Area. Workshop, Great Barrier Reef it is also political because integrated coastal management is Marine Park Authority, Townsville. still not routinely practiced and because different agencies, WALKER, T.A. and O'DONNELL, G.O., 1981. Observations of nitrate, indeed different governments, are responsible for managing phosphate and silicate in Cleveland Bay, northern Queensland. the land on the one hand and the Great Barrier Reef on the Australian Journal Marine Freshwater Research, 32, 877-887. WOLANSKI, E., 1994. Physical Oceanographic Processes ofthe Great other hand. Barrier Reef CRC Press, Boca Raton, 194 pp. WOLANSKI, E. and DUKE,N., 2000. Mud threat to the Great Barrier ACKNOWLEDGEMENTS Reef. In: WANG, Y., and HEALY, T. (Eds.) Muddy Coasts of the World, SCOR publication, Elsevier, in press. This study was supported by the Australian Institute of WOLANSKI, E.; JONES, M., and WILLIAMS, T.J., 1981. Physical prop­ Marine Science. It is a pleasure to thank C. McLean, K. erties of Great Barrier Reeflagoon waters near Townsville. II Sea­ sonal variations. Australian Journal Marine Freshwater Research, Moore, D. Brooks and S. Thomas. 32, 321-334. WOLANSKI, E.; GIBBS, R., RIDD, P., and MEHTA, A., 1992. Settling LITERATURE CITED of ocean-dumped dredged material, Townsville, Australia. Estua­ rine Coastal Shelf Science, 35, 473-490. ANONYMOUS, 1993. The Condition of River Catchments in Queens­ WOLANSKI, E.; SPAGNOL, S., KING,B., and AYUKAI, T., 1999. Patch­ land-a broad overview ofcatchment management issues. Queens­ iness in the Fly River plume in Torres Strait. Journal Marine Sys­ land Department of Primary Industry, Brisbane, 85 pp. tems, 18, 369-381. AYUKAI, T. and WOLANSKI, E., 1997. Importance of biologically me­ YAMAZATO, K., 1987. Effects of deposition and suspension of inor­ diated removal of fine sediments from the Fly River plume, Papua ganic particulate matter on the reef building corals in Okinawa, New Guinea. Estuarine Coastal and Shelf Science, 44,629-639. Japan. Galaxea 6, 289-309.

Journal of Coastal Research, Vol. 16, No.4, 2000